Selection bias in ESG controversies as a risk for sustainable investors

Environmental, Social and Governance (ESG) controversies data, based on the screening of independent media outlets by ESG data providers, has developed into an important dimension of ESG evaluations. Extant research has treated ESG controversies as a black box, applying the dataset without scrutinizing how it is constructed. Our research addresses this knowledge gap, utilizing a two-pronged analytical approach to the ESG controversies data of Vigeo-Eiris. We examine general trends and patterns in the sampling of 20,144 media sources over a 20-year period (2000–2019). This quantitative study is complemented by a qualitative, case-based analysis of the Kazakh operations of a multinational steel manufacturing company to further explore and contextualize the ESG controversies coverage of three leading ESG information providers (Vigeo-Eiris, RepRisk and Sustainalytics). We discover significant selection bias in the media sources that underlie the ESG controversies data, which presents a hitherto undetected risk for investors. Of note, the odds of being covered as part of a controversy are five times higher for companies headquartered in English-language countries than for companies in other language regions. Our study contributes by opening the black box of ESG controversies, scrutinizing their underlying sampling methodologies and highlighting the trade-off between completeness and cost of available ESG information.